A popular question of the day is "what to do about social networks?" A fairly popular and even bipartisan wave says we should "bust Big Tech" (or more tepidly, "regulate" it), but it's unclear what this actually means, or how it would work. The common retort to this sentiment is "a Facebook for different regions of the country (like Bell's breakup) doesn't make sense - we need to be big, because of the networking effect" - the network effect meaning that you and I only use a social network service because other people use it (no one users MySpace, because no one uses MySpace). And if you bust Facebook (the social network) into little Facebooks, "well, it doesn't make sense anymore, and so Facebook (the company) deserves monopoly power over Facebook (the social network)". This argument is however entirely fallacious, and is like stoking fear about pool noodles by pointing to water snakes. We don't need central control for the networking effect to exist.
I want to focus more on this core argument that "Big Tech" wags around, and why it is wrong in three main points. First, we misunderstand how much of Facebook (the social network) can be decomposed, with key parts under monopoly company control, while still preserving the network effect virtually unscathed. Second, we misunderstand the relationship of corporations and the networking effect - it won't be the breakup of Bell, and anyone who tells you that is either misinformed or misleading. Rather, it will be the establishment of a "market of competing services" (I translated to American :P). Elaborating on this, I'll make a quick third point - proprietary social networking software actually impedes the networking effect, by siloing user experience into different services (try talking to someone on Discord from Twitter). The solution is not to just break up Big Tech into smaller monolithic regional companies, nor to give the government control of these. **The solution is to open source and federate.** As a nice side-effect, users can also have more control, security, and privacy. The only compromise required is the powerful in Silicon Valley will lose power.
I will pick on Facebook for the most part, but this applies to any service - Youtube, Twitter, Slack, Discord, etc. Additionally, I will not hit on every single point that can be made, just the big structural ones (sorry, no cool conspiracies :P).
***First***, the public's understanding of what constitutes "Facebook" is frankly poor and inadequate. To most, "Facebook" is a monolithic experience - the network of friends we have on it, the ad services, the style and aesthetic of the website/application, the way our data is harvested and analyzed, and so on. We can conveniently sign up to various services using our Facebook account. Note though that these are component parts of the overall experience - even making these observations indicates that "Facebook" (the experience) is not an irreducible structure. What are the ramifications of this observation?
The core feature of the Facebook experience is the user's account and particular social network. It's imaginable then for Facebook (the company) to publish open-source "connectors" (tecnically called "APIs", think how power outlets standardize electric appliances' power draw, so you can buy any lamp and it will work at your house, even though that lamp designer wasn't involved in your house's construction) which allow software engineers to make their own "Facebook" application. This is done by engineering applications that "connect" into the core Facebook data and software via the API (the connector), but also allow the user to experience Facebook through different styles, different ad experiences, and different algorithms to sort what posts you see - this experience being a choice of service on the part of the user. Reddit is a decent example of this - there are handfuls of different Reddit applications out there, allowing you to have a Reddit experience of your choice. Some are more privacy respecting, some look a little nicer, some look more official. You can even access Reddit through multiple different applications if you want. At the end of the day, all of these applications are "fronts" that connect into the core Reddit structure. You can hopefully start to see how these elements can be decoupled. There is no "networking effect" reason for Facebook (the company) to have a monopoly over ads, application, core structure, and so on.
So which of these elements should Facebook (the company) actually have control over? This is a trick question, as hope will be clarified in the next point. Spoiler: no one should have exclusive control over any of these elements. That a company has exclusive control over all of these elements is quite disturbing.
***Second***, we misunderstand the networking effect - there is no need for central control (corporate or government) for social technology to thrive. The most obvious example of how this works is email. Email is, roughly speaking, an open *protocol* (sort of like how putting things in packages, and writing the destination address in a particular way is a "protocol" for mailing physical things - USPS, FedEx, UPS, everyone shares the basics, at least), and email is something which (for better or worse) we all use. You may have noticed that you can basically email anyone from any account - Hotmail, Outlook, Gmail, Proton Mail, Yahoo, your own email server - it doesn't matter. We are all able to communicate between services without a hitch, and able to choose between services based on various personal preferences. A core protocol shared amongst a diverse array of services has enabled email to remain an extremely easy, common and established way of getting in touch with somebody online. BuT iF yOu HaVe MuLtIpLe SeRvIcEs, YoU lOsE tHe NeTwOrK eFfEcT
If Facebook was "broken up" properly, there wouldn't be "Facebook West Coast" and "Facebook South", and you'd be totally disconnected from your aunt and uncle on "Facebook Midwest", or something like that, which is a ridiculous straw man (at least in the stupid sense). The core protocol would be open and available (like email), and anyone would then be free to develop their own service based on this protocol (like Gmail, Proton Mail, your own email server etc.). The users then get to decide which particular service they want (maybe they prefer the privacy options of one, or perhaps they get a small ad cut from another), while being able to interact with anyone on other services seamlessly because **they're services running the same protocol**. This approach is called **federating**, and is basically opens the floor for a "market of services". In this view, no group would have "control" over any given element, although they would be allowed to compete to provide different suites of advertisement, security, interface, etc.
If Facebook was busted into multiple mini-FBs, the experience for most users would more or less be the same, because each of the mini-FBs would still talk to each other through the shared protocol. If I'm on Facebook-A and I visit a user on Facebook-B, then my service fetches relevant user data from Facebook-B's server (like their profile pic, bio, public photos and posts, etc). If I message that user, then our message logs are shared on our respective servers. I might have ads on my Facebook-A (so I can use it for free), whereas Facebook-B might require an annual fee of a few bucks for server costs, but doesn't have ads.
But besides this "market of services", there is an even more important difference. In the event of a breakup, there would be little initial difference in user experience - in all likelihood, the initial breakup (if done correctly) would be largely unnoticed, as the service (the application) would likely be largely the same, regardless of provider; this would be prudent as an initial demonstration that very little changes for the user experience. *The fundamental difference, in this view, is who controls servers*. As it is now, Facebook (the company) has control over all relevant servers. This gives them enormous power. But in this federated view, we can (1) preserve the networking effect and (2) avoid a central monopoly control over data. Such monopoly control is increasingly leading to issues of censorship and control. For example, the far-right PM of India, Narendha Modhi, is able to blackmail Twitter into shutting down protest accounts by threatening Twitter employees in the country. Or Twitter themselves can unilerally ban people as they please. But the fundamental issue is not the politics of Big Tech themselves (they're powerful capitalists, they'll never be on "our" side), but that centralizing control (power) results in situations where power can, has been, and will be abused. By federating, this problem is rendered virtually moot - imagine trying to ban someone from email, it's a ridiculous idea. This does open the doors for extremist proliferation, but in spite of far-right whining, mainstream social media is still a breeding pool of far-right extremism, therefore I don't see how federating would make things worse (but that's another topic).
There's a ***third*** point as well - closed proprietary social network software actually *hinders* the networking effect. Try to retweet on Facebook. And what inter-social-network integration occurs is due to mergers, which increase the control and power of particular companies, such as Facebook's ownership and integration of Instagram. Is there a way to avoid this, and further, fix this? Yes, it's called **interoperability**. Basically, protocol that bridges between different social media services. This isn't just theory either, it exists as the open-source Matrix protocol (analogous to email), with the flagship service Element (analogous to a specific email provider, ie Proton Mail), and/or opening up interoperability between existing social networks is an idea with bipartisan congressional support via the ACCESS Act.
The most interesting feature of Matrix is that it allows users to communicate with people on various services - Twitter, Discord, etc - from a Matrix service (such as Element). It does this by bridging between these services, and thus breaking down the barriers between these different services. That is to say, open-source software *enhances* the network effect, as it is more amenable to interoperability.
This model - protocol/service decoupling, open-source federation, and interoperability - does not directly solve every problem. For example, in all likelihood, if the government is as lazy with trust-busting as they have been, there will likely still be giants in the social networking world who emerge (similar to how Microsoft and Google are giants in the email world). But it also means that fallacious "monopoly networking effect" argument is not on the table, it does mean it is harder to crack down on dissidence, it does mean more user control. Most importantly, while my faith in the government is quite low, it also means trust-busting of social network companies **makes clear sense** in a way it does not if you take Big Tech's arguments at face value (note that Big Tech has the easiest ability to push propaganda of virtually all industries). Furthermore, it makes social networks less juicy targets for hacks and targeting. Having all data under the control of a single entity not only makes it liable to abuse, it's also liable for targeting. By breaking up the monolith, data breaches do not necessarily have the same scale of damage possible. Note, for example, that Microsoft Outlook's servers recently suffered a huge attack, but that the effect on anyone using other email services was relatively minimal, seeing as most of their data is not on an Exchange server. But if there was one monolithic email provider, we all would have been potentially hit.
Monolithicness is not a price we **have** to pay for the networking effect, in fact it holds us back. We can and should decouple service and protocol (a la email), resulting in the obvious conclusion that multiple competing services (federation) are preferable to a monopolistic, monolithic one.